JP5588302B2 - Medical image processing apparatus and medical image processing method - Google Patents

Description

  The present invention relates to a medical image processing apparatus using a medical image obtained from a medical image imaging apparatus such as an X-ray CT apparatus, an MRI apparatus, an ultrasonic apparatus, and the like, and in particular, efficiently extracts a specific area from image data with high accuracy. It is about the technology to do.

2. Description of the Related Art Conventionally, a region extraction method for extracting individual organs, lesions, and the like (hereinafter referred to as “region of interest”) using medical image data is known. In particular, as a method for automatically detecting a parameter for extracting a region of interest, there is a method disclosed in Patent Document 1.
Japanese Patent Laid-Open No. 2004-228561 uses a technique for expanding a region from a point inside the region of interest (hereinafter referred to as “region expanding method”). The region expansion method starts from an initial region set by an operator or the like, searches for points connected to the initial region from neighboring pixels, extracts a region of interest by capturing the connected points and expanding the region. Is.

In Patent Document 1, a feature amount related to a density value of a pixel in an initial region set in advance from a start point set by an operator is detected, the region is expanded using the detected feature amount, and the expanded region For the density value of the pixel, the feature amount is detected again, and the process of expanding the region according to the newly detected feature amount is repeated, and the extension condition is determined based on the feature amount when the convergence condition is satisfied.
The extraction process using the expansion condition determined by the method of Patent Document 1 is useful when an area having a uniform density value is targeted.

JP 2000-172829 A

  However, the expansion condition determined by the method of Patent Document 1 is constant during the region expansion process. In the region expansion process using a certain expansion condition, for example, a region having a complicated structure such as a contrasted coronary artery region and a variation in density value cannot be extracted with high accuracy. In order to extract such a region with high accuracy, it is necessary to check the state of extraction overflow or extraction omission and to change to appropriate expansion conditions in the region expansion process. Here, “extraction overflow” is a state where pixels outside the region of interest are extracted. In addition, “extraction failure” is a state in which only a region narrower than the region of interest has been extracted.

  The present invention has been made in view of the above-described problems, and the object of the present invention is to perform a complicated operation even when a region having a complex shape and a density value change is set as a region of interest. Another object of the present invention is to provide a medical image processing apparatus and the like that can extract a region with high accuracy and efficiency.

  In order to achieve the above-described object, a first invention is a medical image processing apparatus that extracts a region of interest from medical image data, an image input unit that inputs medical image data to be processed, and the medical image data An extraction start region setting unit for setting an extraction start region, an extraction condition setting unit for setting an extraction condition for extracting a region of interest from the medical image data, the extraction start region as a first generation, and the extraction condition A region extraction processing unit that repeatedly executes a region extraction process, and stores a plurality of generations of extraction results with a set of pixels extracted by one region extraction process as one generation; A determination unit that determines whether or not to continue the region extraction process; and a display unit that displays an image in which the region of interest can be identified based on the extraction result, the extraction condition setting unit , Based on the multiple generations of extraction result, a medical image processing apparatus and updates the extraction conditions.

  A second invention is a medical image processing method for extracting a region of interest from medical image data, an image input step for inputting medical image data to be processed, and an extraction start for setting an extraction start region for the medical image data An area setting step, an extraction condition setting step for setting an extraction condition for extracting a region of interest from the medical image data, and the extraction start area as the first generation, and repeatedly executing the area extraction process based on the extraction condition And a region extraction processing step for storing a plurality of generations of extraction results for a set of pixels extracted by one region extraction processing, and whether to continue the region extraction processing based on the extraction results And a display step for displaying an image in which the region of interest can be identified based on the extraction result. When it is determined that the region extraction processing is to be continued, the processing is repeated from the extraction condition setting step, and the extraction conditions are updated based on the extraction results of the plurality of generations. It is a processing method.

  According to the present invention, a medical image processing apparatus or the like that can efficiently extract a region with high accuracy without complicated operation even when a region having a complex shape and a change in density value is a region of interest. Can be provided.

Hardware configuration diagram of medical image processing apparatus 1 The flowchart which shows the flow of the area | region extraction process in 1st Embodiment Diagram for explaining generations in region extraction processing Diagram showing area expansion in any section Illustration for explaining the extraction overflow criteria The flowchart which shows the flow of the area | region extraction process in 2nd Embodiment Screen display example when extraction overflow occurs Screen display example when the extraction area is discarded Screen display example when changing extraction conditions and restarting extraction

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following description and the accompanying drawings, the same reference numerals are given to the constituent elements having the same functional configuration, and redundant description will be omitted.
First, the configuration of the medical image processing apparatus 1 common to all the embodiments will be described with reference to FIG.

  As shown in FIG. 1, a medical image processing apparatus 1 includes a CPU (Central Processing Unit) 2, a main memory 3, a storage device 4, a display memory 5, a display device 6, a controller 7 to which a mouse 8 is connected, a keyboard 9, The network adapter 10 is connected to the system bus 11 so as to be able to send and receive signals. The medical image processing apparatus 1 is connected to a medical image imaging apparatus 13 and a medical image database 14 via a network 12 so as to be able to send and receive signals. Here, “to enable signal transmission / reception” means a state in which signals can be transmitted / received to each other or from one to the other, regardless of whether they are electrically or optically wired or wireless.

The CPU 2 is a device that controls the operation of each component. The CPU 2 loads a program stored in the storage device 4 and data necessary for program execution into the main memory 3 and executes it. The storage device 4 is a device that stores medical image data captured by the medical image capturing device 13, and is specifically a hard disk or the like. The storage device 4 may be a device that exchanges data with a portable recording medium such as a flexible disk, an optical (magnetic) disk, a ZIP memory, or a USB memory. Medical image data is stored in LAN (Local
Acquired from the medical imaging device 13 and the medical image database 14 via the network 12 such as an area network. The storage device 4 stores a program executed by the CPU 2 and data necessary for program execution. The main memory 3 stores the program executed by the CPU 2 and the progress of arithmetic processing.

The display memory 5 temporarily stores display data to be displayed on the display device 6 such as a liquid crystal display or a CRT (Cathode Ray Tube). The mouse 8 and the keyboard 9 are operation devices for an operator to give an operation instruction to the medical image processing apparatus 1. The mouse 8 may be another pointing device such as a track pad or a track ball. The controller 7 detects the state of the mouse 8, acquires the position of the mouse pointer on the display device 6, and outputs the acquired position information and the like to the CPU 2. The network adapter 10 is for connecting the medical image processing apparatus 1 to a network 12 such as a LAN, a telephone line, or the Internet.
The display device 6, the controller 7, the mouse 8, and the keyboard 9 may be a touch panel display having functions of both a display device and an input device (such as the mouse 8 and the keyboard 9).

The medical image capturing apparatus 13 is an apparatus that captures medical image data such as a tomographic image of a subject. The medical image capturing apparatus 13 is, for example, an MRI apparatus, an X-ray CT apparatus, an ultrasonic diagnostic apparatus, a scintillation camera apparatus, a PET apparatus, or a SPECT apparatus. The medical image database 14 is a database system that stores medical image data captured by the medical image imaging device 13.
When the CPU 2 executes processing to be described later, a region of interest to be observed is extracted, an image in which the region of interest can be identified is created, and the created image is displayed on the display device 6.

The medical image processing apparatus 1 has various functions for extracting a region of interest from medical image data and presenting it to the user through cooperation of various hardware and a program stored in the storage device 4.
Specifically, the medical image processing apparatus 1 extracts a region of interest from medical image data, an image input unit that inputs medical image data to be processed, an extraction start region setting unit that sets an extraction start region for medical image data. An extraction condition setting unit for setting an extraction condition for executing the region extraction processing based on the extraction condition, and a region extraction processing unit for storing the extraction result, and determining whether or not to continue the region extraction processing based on the extraction result And a display unit that displays an image in which the region of interest can be identified based on the extraction result.

<First Embodiment>
The first embodiment will be described with reference to FIGS. 2 to 5 and 7 to 9. In the first embodiment, the CPU 2 (decision unit) of the medical image processing apparatus 1 discards some extraction results, whether to continue the region extraction process, whether to change the extraction condition, or not. Whether or not is determined automatically. Hereinafter, a mode for automatically determining these is referred to as an “automatic determination mode”.

  As shown in FIG. 2, the CPU 2 (image input unit) of the medical image processing apparatus 1 reads medical image data to be processed acquired from the medical image imaging apparatus 13 or the medical image database 14 (step S101).

  Next, the CPU 2 (extraction start area setting unit) analyzes the distribution and shape of the density value of the medical image data to be processed, and sets the extraction start area (step S102). For example, when the region of interest is a coronary artery, the CPU 2 extracts a continuous region having a constant density value in the slice, detects a part of the ascending aorta from the circularity and position information of the continuous region, and extracts the extraction start region. Set. Further, the CPU 2 may set an area designated by the user via the input device (such as the mouse 8 or the keyboard 9) as the extraction start area.

  Next, the CPU 2 (extraction condition setting unit) calculates an extraction condition based on the extraction start area set in step S102 (step S103). In the embodiment of the present invention, the extraction condition calculation processing is not particularly limited. For example, it may be defined by the method described in “JP 2000-172829 A” (Patent Document 1), or may be defined by other known methods.

Next, the CPU 2 (region extraction processing unit) performs region extraction for one generation using the region expansion method from the extraction start region set in step S102 according to the extraction condition calculated in step S103 (step S104). .
The CPU 2 sets the extraction start area as the first generation, repeatedly executes the area extraction process based on the extraction condition, and stores a plurality of generations of extraction results for a set of pixels extracted by one area extraction process. Store in device 4.

FIG. 3 schematically shows the extraction result of each generation when the blood vessel is the region of interest. For example, 20 is the extraction start area (0th generation), 21 is the first generation, 22 is the second generation, 23 is the third generation, and 24 is the fourth generation.
In the graph of FIG. 3, the horizontal axis represents the generation, and the vertical axis represents the number of pixels of each generation (the number of pixels of increments only), indicating a normal blood vessel tendency. As shown in FIG. 3, if the blood vessel is a normal blood vessel, the number of pixels decreases monotonically with each generation if the branching portion is ignored. In the embodiment of the present invention, by utilizing such knowledge, extraction overflow and extraction omission are detected in the region expansion method as described later.

FIG. 4 shows area expansion in any cross section. In step S <b> 104, the CPU 2 searches for pixels that satisfy the extraction condition from neighboring pixels for the previous generation area and extracts them as a new generation area.
In FIG. 4, when N is a generation in which attention is paid to N (N is a positive integer), 25 is the (N-2) th generation, 26 is the (N-1) th generation, and 27 is the Nth generation. The (N-2) generation includes four pixels. The (N-1) th generation includes all adjacent pixels for the (N-2) th generation and includes 8 pixels. The Nth generation includes a part of adjacent pixels to the (N−1) th and includes nine pixels.

  In the above description, in step S104, the CPU 2 performs region extraction for one generation. However, region extraction for a plurality of generations may be performed collectively. That is, the CPU 2 may proceed to step S105 after extracting regions for a plurality of generations (for example, for three generations) in step S104. As a result, the number of times step S105 to step S107 and step S109 to step S112, which will be described later, are executed can be reduced, and the overall processing can be speeded up.

Returning to the description of FIG.
Next, the CPU 2 (determination unit) determines whether or not to determine whether to stop extraction based on the extraction result in step S104 (step S105). Specifically, the CPU 2 compares the number of extracted pixels of the extraction result (the extraction result of the generation of interest) in step S104 with a threshold value, and if it is equal to or less than the threshold value (YES in step S105), the process proceeds to step S106. . On the other hand, if it is greater than the threshold value (NO in step S105), the CPU 2 determines to continue the region extraction process and proceeds to step S109.
Here, the threshold value is “0”, for example. In this case, the CPU 2 proceeds to step S106 if the extracted result of the generation of interest is 0, and proceeds to step S109 if it is not 0.

  If the CPU 2 collectively extracts a plurality of generations in step S104, in step S105, the CPU 2 sets the number of extracted pixels and the threshold value of the extraction result of the last generation stored in the storage device 10. Compare.

In step S106, the CPU 2 (determination unit) determines whether or not an extraction omission has occurred. Specifically, the CPU 2 determines whether or not a preset termination condition is satisfied.
When determining that the end condition is satisfied (NO in step S106), the CPU 2 (determination unit) determines to end the region extraction process. Then, the CPU 2 (display unit) sets image data in which the region of interest can be identified based on the extraction result in the display memory 5 and displays it on the display device 6 (step S108).
If the CPU 2 (determination unit) determines that the termination condition is not satisfied (YES in step S106), it determines that an extraction omission has occurred. Then, the CPU 2 (extraction condition setting unit) relaxes (changes) the extraction condition (step S107). The degree of relaxation of the extraction condition may be set in advance with the step size of the extraction condition parameter. Here, in order to simplify the processing (in order to increase the speed), the CPU 2 relaxes the extraction condition using the set step size, but the number before it is determined that the extraction omission has occurred. The extraction condition may be set by analyzing the average density value of the generation, or the extraction condition may be determined by analyzing the pixels around the pixel extracted immediately before the extraction omission occurs.

Here, preset termination conditions will be described.
The first end condition is, for example, that the global parameter of the extraction condition illustrated in 38 (38a to 38c) in FIGS. 7 to 9 reaches a preset upper limit value or lower limit value. Here, the global parameters are the maximum value and the minimum value of the density values in the extraction region. Therefore, the first end condition is that the maximum value of the density value in the extraction region has reached the preset upper limit value or the lower limit for which the minimum value is preset as the global parameter of the extraction condition Either the value has been reached. 7 to 9 will be described later.
The second end condition is that, in step S105, it is determined that the number of extracted pixels of the extraction result of the generation of interest is equal to or less than a threshold value continuously for a predetermined number of times.
The third end condition is that a specific pixel set in advance is extracted. The specific pixel may be designated by the user via an input device or the like, or may be automatically set by the CPU 2. In this case, the specific pixel serves as a boundary between the region of interest and the extraction prohibited region.
In the embodiment of the present invention, any of the first to third end conditions may be used, a plurality of conditions may be used in combination, or other end conditions may be used.

Returning to the description of FIG.
When the process proceeds to step S109, the CPU 2 has determined to continue the area extraction process. In step S109, the CPU 2 (determination unit) determines whether or not an extraction overflow has occurred based on a change in the number of extracted pixels as an extraction result.

Here, with reference to FIG. 5, a determination condition (extraction overflow determination condition) relating to a process for determining whether or not an extraction overflow has occurred will be described.
In FIG. 5A, the horizontal axis represents the extraction process (generation) and the vertical axis represents the number of pixels, and the change in the number of pixels is shown. As shown in FIG. 5 (a), during the extraction process, the transition of the number of extracted pixels is observed for each generation, and if there is a sudden change or the like different from the previous transition, it can be considered that the extraction overflow has occurred. .
In order to find an abrupt change, the CPU 2 calculates, for example, a moving average of the number of extracted pixels of the extraction results of a plurality of generations (for example, three generations), the latest moving average value, and the moving average value before the latest. If the difference between the two is greater than or equal to the threshold, it is determined that extraction overflow has occurred. Compared with a method for determining whether or not an extraction overflow has occurred based on an extraction result of one generation, the method according to the embodiment of the present invention significantly improves the determination accuracy of an extraction overflow.
The extraction condition can be updated at an appropriate timing. As a result, even when a region having a complex shape and a change in density value is set as a region of interest, the region is extracted with high accuracy and efficiency. be able to.

In FIG. 5B, the horizontal axis represents the extraction process (generation) and the vertical axis represents the average density value, and the change in the average density value for each generation is shown. As shown in FIG. 5 (b), the transition of the average density value is observed for each generation in the extraction process, and if there is a sudden change or the like different from the transition so far, it can be considered that the extraction overflow has occurred. . Here, the determination condition for the occurrence of extraction overflow does not have to be a “rapid” change in the average density value. For example, the CPU 2 may determine that an extraction overflow has occurred when the density value falls below or exceeds a preset density value.
Therefore, the CPU 2 may determine whether or not an extraction overflow has occurred based on a change in the average density value of the extraction result in addition to a change in the number of extracted pixels as the extraction result.
For example, the CPU 2 calculates the moving average of the average density values of the extraction results of a plurality of generations (for example, three generations), and the difference between the most recent moving average value and the moving average value before the latest is equal to or greater than the threshold value Alternatively, it may be determined that an extraction overflow has occurred.
Further, for example, instead of the moving average, the CPU 2 calculates a differential value of the average density value of the extraction result, and determines that an extraction overflow has occurred when the absolute value of the differential value is equal to or greater than a threshold value. Also good.

For example, in the extraction of a coronary artery region, the transition of the number of pixels may change differently from that in a portion where a branching portion or an aneurysm exists. In this case, it is desired to extract the bifurcation and aneurysm as the coronary artery region, but if the CPU 2 determines whether or not the extraction overflow has occurred based only on the change in the number of pixels, the bifurcation or aneurysm is regarded as the extraction overflow, A bifurcation or an aneurysm is no longer extracted as a region of interest. As described above, there is a case where it is impossible to distinguish whether the extraction overflows to an unintended region or the region of interest is correctly extracted only by changing the number of pixels.
Even in such a case, the extraction overflow determination accuracy is improved by adding a change in the average density value of the extraction result as the extraction overflow determination condition.

For example, in the extraction of the coronary artery region, if the overflow has spread to the myocardial pixels, the average density value can be considered to be a lower value. By adding, it becomes possible to distinguish the coronary artery region (including a portion where a bifurcation or an aneurysm is present) from the myocardium.
Further, for example, in the extraction of the coronary artery region, when the coronary artery region and the bone region are in contact with each other, and the overflow has spread to the pixels of the bone region, it is possible that the average density value shows a higher value. By adding a change in the average density value of the extraction result to the determination condition, it is possible to distinguish the coronary artery region from the bone region.

In addition to the above-described extraction overflow determination conditions, it may be determined whether or not an extraction overflow has occurred by analyzing the shape of the extraction region in the extraction process.
For example, in the blood vessel extraction, the CPU 2 may determine whether or not an extraction overflow has occurred based on a change in the circularity of the cross section of the blood vessel. In this case, the CPU 2 determines that an extraction overflow has occurred when the circularity of the cross section of the blood vessel changes rapidly.
In addition, for example, in the blood vessel extraction, the CPU 2 calculates the barycentric coordinates for each generation, and the generations having consecutive generations (for example, the 0th generation and the 1st generation, the 1st generation and the 2nd generation, the 2nd generation) And the third generation,...) May be calculated, and a determination may be made as to whether or not extraction overflow has occurred based on a change in the direction of the vector. In this case, the CPU 2 determines that an extraction overflow has occurred when the direction of the vector changes abruptly.
These are considered useful in extracting a region having a linear structure with little change, such as a lower limb blood vessel.

Returning to the description of FIG.
If the CPU 2 (determination unit) determines that an extraction overflow has occurred (YES in step S109), the CPU 2 proceeds to step S110, and if it determines that an extraction overflow has not occurred (NO in step S109). ), The process is repeated from step S104.

When the process proceeds to step S110, it is determined that the extraction overflow has occurred. In step S110, the CPU 2 (decision unit) decides to change the extraction condition and decides to discard the extraction results of some generations (generations where extraction overflow occurs).
Then, the CPU 2 (extraction condition setting unit) tightens (changes) the extraction condition (step S110), and the CPU 2 (extraction start area setting unit) discards the extraction result of the generation in which the extraction overflow occurs ( In step S111), extraction results up to generations that have not been discarded are reset as extraction start areas (step S112), and the processing is repeated from step S104. In order to make the extraction conditions stricter, the step size of the extraction condition parameters may be set in advance. Here, in order to simplify the processing (in order to increase the speed), the CPU 2 uses a set step size to tighten the extraction conditions, but before the determination that an extraction overflow has occurred has occurred. An extraction condition may be set by analyzing an average density value of several generations, or an extraction condition may be determined by analyzing a pixel around a pixel extracted immediately before an extraction overflow occurs.

  In FIG. 2, for easy understanding, step S111 and step S112 are shown in two steps, but the internal processing actually performed by the CPU 2 needs to be considered separately. Absent. That is, the CPU 2 releases the memory area of the extraction result of the generation in which the extraction overflow has occurred (discards the extraction result of the generation in which the extraction overflow has occurred). Among them, the extraction result of the latest generation is reset as the extraction start area. For example, when the CPU 2 discards the extraction result of N that is the last extracted generation, the extraction result of N-1 that is the previous generation is newly set as the extraction start area.

If the CPU 2 collectively extracts a plurality of generations in step S104, in step S111, the CPU 2 selects an extraction start region from a plurality of generation extraction results based on a plurality of region extraction processes. Specify the extraction results of the generation that is the standard for resetting.
For example, when the region extraction for three generations (N-2, N-1, and N) is collectively performed and the generations where extraction overflow occurs are N-1 and N, the CPU 2 The extraction results of the N-1 and N generations are discarded, and the extraction result of the N-2 generation is reset as the extraction start area.

  As described above, even when the medical image processing apparatus 1 according to the first embodiment has a complex shape and a region having a change in density value as a region of interest, the region can be efficiently and efficiently without complicated operations. Can be extracted.

<Second Embodiment>
Next, a second embodiment will be described with reference to FIGS. In the second embodiment, the CPU 2 (decision unit) of the medical image processing apparatus 1 determines whether or not to continue the region extraction process based on the instruction data input by the user via the input device or the like. It is manually determined whether or not to change and whether to discard some extraction results. Hereinafter, a mode for manually determining these is referred to as a “manual determination mode”.

  As shown in FIG. 6, the CPU 2 (image input unit) of the medical image processing apparatus 1 reads medical image data to be processed acquired from the medical image imaging apparatus 13 or the medical image database 14 (step S201).

Next, the CPU 2 (display unit) sets the image data on which the extraction region is not superimposed, that is, the medical image data to be processed that is read in step S201, in the display memory 5 and displays it on the display device 6 (step S202). ).
For example, the CPU 2 displays medical image data to be processed as 31 (31a to 31c), 32 (32a to 32c), and 33 (33a to 33c) illustrated in FIGS. Further, the CPU 2 may display the three-dimensional image data to be processed (before the extraction area is reflected) as 34 (34a to 34c) illustrated in FIGS. 7 to 9 will be described later.

  Next, the CPU 2 (extraction start area setting unit) reads the extraction start area data acquired from the medical image imaging device 13 or the medical image database 14 (step S203). Further, the CPU 2 may read an area designated by the user via the input device (such as the mouse 8 and the keyboard 9) as extraction start area data.

  Next, the CPU 2 (extraction condition setting unit) sets the extraction condition based on the extraction start area data read in step S203 (step S204). Further, the CPU 2 may set a parameter value designated by the user via the input device (such as the mouse 8 or the keyboard 9) as the extraction condition.

Next, the CPU 2 (region extraction processing unit) performs region extraction for one generation using the region expansion method from the extraction start region data read in step S202 according to the extraction condition set in step S204 (step S205). .
The CPU 2 sets the extraction start area as the first generation, repeatedly executes the area extraction process based on the extraction condition, and stores a plurality of generations of extraction results for a set of pixels extracted by one area extraction process. Store in device 4.
Although the CPU 2 performs the region extraction for one generation, the region extraction for a plurality of generations may be collectively performed as in the first embodiment.

Next, the CPU 2 (display unit) displays the extraction result obtained by the region extraction process in step S205 on the display device 6 (step S206).
Then, the CPU 2 (determination unit) determines whether or not to determine whether to stop extraction based on instruction data input by the user via an input device or the like (step S207). In step S207, the CPU 2 may determine whether to automatically determine whether to stop extraction, as in the first embodiment.
If the CPU 2 (decision unit) decides to make an extraction stop decision (YES in step S207), it proceeds to step S208, and if it decides not to make an extraction stop decision (NO in step S207), from step S205 Repeat the process.

In step S208, the CPU 2 (determination unit) determines whether or not to change the extraction condition based on instruction data input by the user via the input device or the like (step S207).
When the CPU 2 (decision unit) decides to change the extraction condition (YES in step S208), the CPU 2 (extraction condition setting unit), based on the instruction data input by the user via the input device or the like, The extraction condition is changed (step S209), and the process proceeds to step S210.
If the CPU 2 (determination unit) determines not to change the extraction condition (NO in step S208), the process proceeds to step S210.

FIG. 7 is an example of a screen display when the region extraction processing is performed using the region of interest as the coronary artery and extraction overflow occurs.
Reference numerals 31a to 33a denote cross-sectional images to be processed (shaded portions are extraction regions). 31a to 33a are, for example, images of three orthogonal cross sections of axial, coronal, and sagittal. Moreover, 31a-33a is good also as an MPR image of arbitrary cross sections.
Reference numeral 34a denotes a three-dimensional image to be processed (the hatched portion is an extraction region).

35a is a graph showing a change in the number of pixels for each generation of extraction results.
36a is a graph showing a change in average density value for each generation of extraction results. Two dotted lines extending in the left-right direction in 36a indicate general parameters of the extraction condition, that is, the maximum value and the minimum value of the density value.

Reference numeral 37a denotes a scroll bar for discarding the extraction result for each generation. One dotted line extending vertically in 36a moves in conjunction with the scroll bar 37a. The user slides the scroll bar 37a so that the generation to be discarded is positioned on the right side of the dotted line. And CPU2 determines the generation which discards an extraction result based on the instruction | indication by a user.
That is, the CPU 2 displays an object indicating the position of the generation overlaid on the graph representing the change in the number of pixels and the average density value, and determines the generation for discarding the extraction result based on the position of the object.

  Reference numeral 38a denotes an operation panel for setting global parameters and local parameters of extraction conditions. As described above, the global parameters of the extraction condition are the maximum value and the minimum value of the density values in the extraction region. The global parameter is a reference value for determining a global change in the extraction region. The local parameter of the extraction condition is a value obtained by applying a weight to the average value (average error) of errors between adjacent pixels in the extraction region. The local parameter is a reference value for determining a local change in the extraction region.

  In the operation panel 38a, a value can be directly specified in a numerical value input box ("Upper", "Lower", "+", "-"). Further, on the operation panel 38a, by selecting the “strict” button and the “relax” button, the CPU 2 can automatically calculate a value as in the first embodiment.

For example, when the user selects the “strict” button using the input device, the CPU 2 sets the global parameters to “Upper−α” and “Lower + α” based on a preset parameter step α (α> 0). To "".
Further, for example, when the user selects the “relax” button by the input device, the CPU 2 sets the global parameters to “Upper + β”, “Lower−” based on a preset parameter step β (β> 0). Set to “β”. Further, as described above, the CPU 2 may set the extraction condition by analyzing the average density value of several generations before it is determined that the extraction omission or the extraction overflow has occurred, or the extraction omission or the extraction overflow. Alternatively, the extraction condition may be determined by analyzing pixels around the pixel extracted immediately before the occurrence.

Further, when the user selects the “strict” button or the “relax” button by the input device, the CPU 2 may change only one of the values “Upper” and “Lower”.
Further, when the user selects the “strict” button using the input device, the CPU 2 may decrease the values of the local parameters “+” and “−”. Further, when the user selects the “relax” button using the input device, the CPU 2 may increase the values of the local parameters “+” and “−”.

  Reference numeral 39 denotes a button for instructing stop / restart of the extraction process. Reference numeral 40 denotes a button for ending extraction.

As illustrated in FIG. 7, when the CPU 2 finishes the region extraction process in step S205, the CPU 2 displays a graph indicating changes in the number of extracted pixels and average density values as extraction results. As a result, as in the automatic determination mode in the first embodiment, the user can judge more accurately when compared with a case where it is determined whether or not an extraction overflow or the like has occurred by referring to only one generation extraction result. Greatly improved.
The extraction condition can be updated at an appropriate timing. As a result, even when a region having a complex shape and a change in density value is set as a region of interest, the region is extracted with high accuracy and efficiency. be able to.

As can be seen from 34 a in FIG. 7, the extraction overflow spreads from the coronary artery region to the myocardial pixels. This can be grasped quantitatively and clearly from the fact that the number of pixels is rapidly increasing at 35a, and the average density value has rapidly decreased at 36a and has not changed to a small value. be able to.
Therefore, the user instructs the change of the extraction condition by directly changing the value of the global parameter or the local parameter or selecting the “strict” button via the input device.

Returning to the description of FIG.
In step S210, the CPU 2 (determination unit) determines whether or not to discard the extraction area based on instruction data input by the user via an input device or the like.
When the CPU 2 (determination unit) determines to discard the extraction area (YES in step S210), the CPU 2 (extraction area setting unit), based on instruction data input by the user via an input device or the like, The extraction area is discarded (step S211), the extraction results up to the generation that has not been discarded are reset as the extraction start area (step S212), and the process proceeds to step S213.
When the CPU 2 (determination unit) determines not to discard the extraction area (NO in step S210), the process proceeds to step S213.

In step S213, the CPU 2 (decision unit) performs extraction based on instruction data input by the user via an input device or the like, that is, based on whether or not an instruction to change the extraction condition or discard the extraction area is given. Determine whether to end the process.
When the CPU 2 (determination unit) determines to end the extraction process (YES in step S213), the process ends.
When the CPU 2 (determination unit) determines not to end the extraction process (NO in step S213), the process is repeated from step S205.

FIG. 8 shows a screen display example when the extraction area is discarded.
The user slides the scroll bar 37a shown in FIG. 7 to the left until reaching the position of the scroll bar 37b shown in FIG. 8, and instructs the discard of the extraction area.
In the example shown in FIG. 8, the user instructs to discard the extraction area until the generation in which the number of pixels is rapidly increasing in 35b or the generation in which the average density value is rapidly decreased in 36b.
In each of 31b to 34b, the discarded generation extraction areas are not displayed. Thereby, the user can confirm the discard result of the extraction area in real time.
In 35b and 36b, the position of the object indicating the generation position (one dotted line extending in the vertical direction) is changed.
In 38b, neither the global parameter nor the local parameter is changed.

FIG. 9 is a screen display example when the extraction condition is changed, the extraction process is restarted, and a new extraction result is reflected.
The user tightens (changes) the value of “Lower” of the global parameter on the operation panel from “130” of 38b shown in FIG. 8 to “200” of 38c shown in FIG.
Then, the CPU 2 performs region extraction processing based on the changed extraction condition, and displays the extraction results on 31c to 36c.

As can be seen from 34c in FIG. 9, there is no overflow from the coronary artery region to the myocardial pixels, and the coronary artery region is accurately extracted to the extreme end.
This can be grasped quantitatively and clearly because there is no portion where the number of pixels is rapidly increasing in 35c and there is no portion where the average density value is rapidly decreasing in 36c. .
Thus, displaying the graph of the change in the number of pixels and the average density value on the display device 6 is for determining whether or not the extracted region of interest (in this case, the coronary artery region) is accurately extracted. It also presents quantitative judgment materials.

  As described above, the medical image processing apparatus 1 according to the second embodiment can efficiently and accurately perform a region without complicated operations even when a region having a complex shape and a density value change is a region of interest. Can be extracted.

  The preferred embodiments of the medical image processing apparatus and the like according to the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples. It will be apparent to those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea disclosed in the present application, and these naturally belong to the technical scope of the present invention. Understood.

1 ... Medical image processing device 2 ... CPU
3 ......... Main memory 4 ......... Storage device 5 ......... Display memory 6 ......... Display device 7 ......... Controller 8 ......... Mouse 9 ......... Keyboard 10 ......... Network adapter 11 ......... System Bus 12 ……… Network 13 ……… Medical image capture device 14 ……… Medical image database

Claims (14)

A medical image processing apparatus for extracting a region of interest from medical image data,
An image input unit for inputting medical image data to be processed;
An extraction start area setting unit for setting an extraction start area of the medical image data;
An extraction condition setting unit for setting an extraction condition for extracting a region of interest from the medical image data;
The extraction start area is the first generation, the area extraction process is repeatedly executed based on the extraction conditions, and a plurality of generations of extraction results are stored with one set of pixels extracted by one area extraction process. A region extraction processing unit to perform,
A determination unit that determines whether to continue the region extraction process based on the extraction result;
A display unit that displays an image in which the region of interest can be identified based on the extraction result;
With
The extraction condition setting unit updates the extraction condition based on the plurality of generation extraction results. When the determination unit determines to continue the region extraction process, the determination unit determines whether to change the extraction condition ;
Furthermore, when the determination unit determines to change the extraction condition and discard the extraction results of some generations, the extraction start area setting unit extracts the generation immediately preceding the generation to be discarded The medical image processing apparatus according to claim 1, wherein a result is reset as the extraction start area. When the determination unit is in the automatic determination mode,
The determination unit determines to continue the region extraction process when the number of extracted pixels of the extraction result of the generation of interest is greater than a threshold, and based on a change in the number of extracted pixels of the extraction result, Determine whether or not
Further, when the determination unit determines that an extraction overflow has occurred, the determination unit determines to change the extraction condition, determines to discard the extraction results of some generations, and the extraction condition setting unit 3. The medical image processing apparatus according to claim 2, wherein the extraction condition is changed, and the extraction start area setting unit discards an extraction result of a generation in which extraction overflow occurs. When the determination unit is in the manual determination mode,
When the region extraction processing unit finishes the region extraction processing, the display unit displays a change in the number of extracted pixels of the extraction result as a graph,
3. The medical image according to claim 2, wherein the determination unit determines to continue the region extraction process, to change the extraction condition, and to discard the extraction result based on input instruction data. Processing equipment.   The determination unit determines to continue the region extraction process, change the extraction condition, and discard the extraction result when the region extraction processing unit finishes the region extraction process a plurality of times. The medical image processing apparatus according to claim 2. The determination unit determines whether or not a preset end condition is satisfied when the number of extracted pixels of the extraction result of the last generation stored in the region extraction processing unit is equal to or less than a threshold value,
Further, when the determination unit determines that the end condition is satisfied, the determination unit determines to end the region extraction process,
Furthermore, when the determination unit determines that the termination condition is not satisfied, the determination unit determines that an extraction omission has occurred, and the extraction condition setting unit changes the extraction condition. The medical image processing apparatus according to claim 3.   The determination unit calculates a moving average of the number of extracted pixels of the extraction results of a plurality of generations, and if the difference between the most recent moving average value and the moving average value before the most recent is equal to or greater than a threshold value, extraction overflow occurs. The medical image processing apparatus according to claim 3, wherein the medical image processing apparatus is determined to have occurred.   The medical image processing apparatus according to claim 3, wherein the determination unit further determines whether extraction overflow has occurred based on a change in an average density value of the extraction result.   The medical image processing apparatus according to claim 3, wherein the determination unit further determines whether or not an extraction overflow has occurred based on a change in circularity of a cross section of the extraction result.   The determination unit further calculates a center-of-gravity coordinate for each generation, calculates a vector defined by the center-of-gravity coordinate between generations with successive generations, and an overflow of extraction occurs based on a change in the direction of the vector The medical image processing apparatus according to claim 3, wherein it is determined whether or not it is performed.   The medical image processing apparatus according to claim 4, wherein the display unit further displays a graph of a change in average density value of the extraction result when the region extraction processing unit finishes the region extraction processing. The display unit displays an object indicating the position of the generation, superimposed on the graph display,
The medical image processing apparatus according to claim 4, wherein the determination unit determines a generation for discarding the extraction result based on the position of the object.   When the determination unit determines to discard the extraction results of some generations, the extraction start region setting unit starts the extraction from a plurality of generation extraction results based on the plurality of region extraction processes. The medical image processing apparatus according to claim 5, wherein the extraction result of the generation reset to the area is specified. A medical image processing method for extracting a region of interest from medical image data,
An image input step for inputting medical image data to be processed;
An extraction start area setting step for setting an extraction start area of the medical image data;
An extraction condition setting step for setting an extraction condition for extracting a region of interest from the medical image data;
The extraction start area is the first generation, the area extraction process is repeatedly executed based on the extraction conditions, and a plurality of generations of extraction results are stored with one set of pixels extracted by one area extraction process. Region extraction processing step to perform,
A determination step for determining whether or not to continue the region extraction process based on the extraction result;
A display step of displaying an image in which the region of interest can be identified based on the extraction result;
With
When the determination step determines to continue the region extraction process, the process is repeated from the extraction condition setting step, and the extraction condition is updated based on the extraction results of the plurality of generations. Image processing method.

Images